What FCC catalyst tests show

The significance of equilibrium catalyst tests are reviewed using results from typical catalyst reports. These reports can help identify operating trends
L.L. Upson, Katalistiks b.v., Surte, Sweden ALL CRACKING CATALYST manufacturers routinely test equilibrium FCC catalyst samples from their customers. It is the purpose of this paper to examine the types of data typically found in an equilibrium catalyst test report, to describe briefly how this information is obtained and what the significance of these data are to the refiner. The equilibrium catalyst report is one of the ways catalyst manufacturers and refiners can work together to optimize the operations of cat crackers. Each item in the report can provide valuable information to help better understand the operations of the cat cracker. Better understanding then leads to better operations. Catalyst reports are most useful as a trend indicator. When changes in catalyst properties are detected, the FCC management is alerted to look for the cause of the change and if necessary to modify the operations to return to more normal conditions. In some cases the refiner will be able to get two or more catalyst makers to test his equilibrium catalyst. In such a case, the refiner should not expect the absolute values from the two laboratories to agree they probably wont. In many instances the values reported depend upon the testing techniques used, and there is no standardization of techniques among the catalyst makers. What the refiner should look for is that the two reports show the same trends. Since all catalyst data have some variability associated with them (due both to sampling variability and to testing variability) a small change from one test period to another can represent a trend or it can merely be due to normal variability. If the results from both laboratories show the same trend, it is very likely that the reported change represents a real shift in catalyst properties. A typical report is shown in Fig. 1. The results concern catalytic properties (MAT results), physical properties and chemical analysis. A description of the tests among these categories follows: CATALYTIC PROPERTIES Catalytic properties are determined by a micro-activity test (MAT). A sample of the decoked equilibrium catalyst is placed in a laboratory reactor, a typical FCC feed is passed over the catalyst at cracking conditions and the results analyzed. The MAT test conditions are listed in Table 1. Activity. The value reported as activity is the conversion achieved in the MAT laboratory test. Since this conversion is obtained on a decoked sample at standard test conditions, any change in activity is due to catalyst changes. In commercial operations it is sometimes difficult to determine the reason for a change in unit conversion. It can be due to changes in processing conditions, changes in feed quality, or changes in catalyst characteristics. The MAT activity data allows the refiner to separate the catalyst effect from the process and feed effect when he evaluates unit performance. To minimize the variability in the MAT value, we run each sample in duplicate and sometimes in triplicate.

KATALISTIKS EQUILIBRIUM CATALYST TEST REPORT COMPANY: _______________________

MAT Results ID No. Date Taken Act. CF GF H2/CH4 0-20 Particle Size Distribution (micr.) 0-40 0-60 0-80 0-100 Avg. Surf Area m /g 268 1981 23/1 Fig. 1 Example of an equilibrium catalyst report. 61 1.32 1.14 0.9 0.4 5.8 30 61 79 72 64
2

Physic. Properties Pore Vol. cc/g 0.25 Bulk dens. G/cc 0.85 180 12 Ni ppm Cu ppm

Chemical Analysis V ppm Fe Wt% Na2O Wt% C Wt% Remarks

880

0.33

39

0.17

We believe that the variability of the MAT results from the Katalistiks laboratory is as low or lower than any other catalyst laboratory. Nevertheless, the 95% confidence level of our MAT activity value is still 1 numbers. Thus it always takes at least two consecutive samples to establish confidence that the activity changes represent a real trend. A change in catalyst activity should produce a corresponding change in the conversion of the commercial unit, providing such things as reactor temperature and feed quality remain constant. Normally, unit conversion will not change as much as the MAT activity value changes. As a general rule, the expected change in unit conversion will be about 80-90% of the change in MAT activity. Yet many factors can influence the MAT activity value. Catalyst makeup rate is important. For a typical unit with an average catalyst makeup rate of 1% per day, doubling the catalyst makeup rate would be expected to increase the MAT activity value 4-6 numbers. Feed metals contamination reduced catalyst activity. A typical example of the effect of metals is shown in Table 2. A total metals level of less than 1000 ppm has very little effect upon equilibrium catalyst activity. Activity then begins to drop off rapidly as the metals level exceeds 2000 ppm. Catalyst poisoning by other contaminants, particularly sodium, can cause significant loss in activity. Sodium acts as a fluxing agent for alumina, reducing the melting point of the catalyst structure. Normal regenerator temperatures are then sufficiently high to melt the contaminated part of the catalyst, destroying both sieve and matrix. When considering sodium content in equilibrium catalyst, it is necessary to relate it to the sodium content of fresh catalyst, since it is the sodium between equilibrium and fresh catalyst that is important. Sodium content of fresh zeolitic catalysts depends mainly on the sieve content of the catalyst. There is thus a wide difference in sodium content of properly prepared fresh catalyst, with some containing as little as 0.3 wt. % and some as high as 1.1 wt. %. The sodium in fresh catalyst is associated with the zeolite and is buried deep in the zeolite cage. This sodium is essentially inert with respect to fluxing with alumina. It is the sodium coming with the cat cracker feed that presents a catalyst deactivation problem. The increase in sodium content above the normal fresh catalyst level thus gives an indication of sodium poisoning due to cracker feed sodium contamination. Until the sodium poisoning problem is solved, the refiner will usually increase fresh catalyst makeup rate to

maintain a constant activity in his circulating catalyst inventory. Regeneration can reduce catalyst activity by the combined action of steam and high temperatures. This process is known as hydrothermal deactivation. At normal regenerator conditions the catalyst is exposed to a steam atmosphere equivalent to 160-200 mm Hg water partial pressure. At this water concentration, with regenerator temperatures in the region of 650-700C, a substantial part of the matrix (40-60%) and a smaller portion of the molecular sieve (25%) is quickly destroyed after only a few cycles through the regenerator. The remaining catalyst is very stable. Essentially, no further destruction occurs to the matrix1 while the molecular sieve portion deactivates slowly at the rate of less than 0.5% per day. At regenerator temperatures in the range of 650-700C, this hydrothermal deactivation is essentially independent of temperature.
TABLE 1 Micro activity test (MAT) conditions
Catalyst quantity, g Feed quantity, g Cat/oil ratio Reaction time, sec Space velocity (WHSV), hr Reactor temperature, C
-1

4.0 1.3 3.0 75 15 482 (900F)

Feed (ASTM standard feed) Boiling Point Initial 5 vol% 50 vol% 90 vol% Final Sulfur, wt% Rams carbon, wt% UOP K

198 262 412 552 572 0.64 0.21 12.0

388 504 773 1025 1061

TABLE 2 Equilibrium catalyst, metals vs. activity

Ni + V, Total ppm 440 2500 7000 Activity, MAT 70 68 60

At higher regenerator temperatures, hydrothermal destruction of the molecular sieve increases rapidly with a corresponding loss of equilibrium catalyst activity. It has been reported2 that increasing regenerator temperature from 700C to 725C can sufficiently deactivate a typical commercial cracking catalyst to cause a 4-6 number drop

in MAT activity. This activity loss then requires that the fresh catalyst makeup rate be doubled to maintain the same system activity. Exposure to very high temperatures can be particularly disastrous. Even a short exposure can cause substantial deactivation. As an example, in one unit which had a regenerator temperature excursion to 900C for a few minutes followed by a one hour period at 800C before dropping back to normal conditions below 700C, the MAT activity of the equilibrium catalyst decreased 6 numbers. This would indicate that approximately 25% of the zeolite activity in the catalyst had been destroyed in less than three hours. Increased the water concentration in the regenerator, via the use of excessive quench steam, greatly increases the rate of hydrothermal deactivation. For example, increasing the water concentration from the normal regenerator level of 180 mm Hg to 300 mm Hg partial pressure at mild regenerator temperatures (700C) would results in a 2-3 number drop in activity1 or a 30-50% increase in catalyst addition to maintain constant activity. Catalyst Selectivity. In addition to providing information on catalytic activity, the equilibrium catalyst report also gives information on catalytic selectivity. On the Katalistiks report this information is shown in the form of a Coke Factor (CF), a Gas Factor (GF), and a H2/GH4 molar ratio. The Coke Factor (CF) describes the coke-making characteristics of the equilibrium catalyst relative to the coke-making characteristics of some standard catalyst at the same conversion. In our laboratory the coke vs. conversion relationship of steam-aged EKZ catalyst is used as the reference curve. The Coke Factor is influenced by the type of fresh catalyst used and is strongly influenced by the metals in the feed that are deposited upon the equilibrium catalyst. Coke yields respond almost instantly to metals deposition. Total metals (Ni + V) as low as 500 ppm can produce CF values 10-30% above fresh catalyst values. It has been reported3 that an increase of 2000 ppm of nickel plus vanadium in the equilibrium catalyst from a typical feed at typical catalyst makeup rate (1% of unit inventory) would double the Coke Factor of the equilibrium catalyst. The Coke Factor relates directly to the temperature in the regenerator. The temperature in the regenerator is influenced by the feed preheat temperature, the reactor temperature, the carbon residue of the feed and the cokemaking properties of the equilibrium catalyst. The Coke Factor isolates the catalyst contribution, which is the combination of the fresh catalyst plus the accumulated metals. The Gas Factor (GF) describes the light hydrocarbon production (C1-C4) of the equilibrium catalyst relative to

the C1-C4 production of a standard catalyst at the same conversion. The technique for determining a GF is the same as is used in determining a CF. The gas yield of the equilibrium catalyst is compared to the gas yield of the standard catalyst. At Katalistiks we again use steam-aged fresh EKZ catalyst as our standard catalyst. The Gas Factor , like the Coke Factor, is a function of the type of fresh catalyst used and of the contaminant metals deposited on the catalyst. The GF, however, is much less sensitive to metals than the CF. Total metals (Ni+V) of 1500-2000 ppm are required before the GF begins to increase above fresh catalyst values. At metal levels normally found in todays equilibrium catalysts the Gas Factor is more an indicator of the type of catalysts the Gas Factor is more an indicator of the type of catalyst used than it is of metals contaminants. The Gas Factor is useful to evaluate FCCU gas plant operating problems (gas compressor, absorber, stripper, etc.). The gas plant capacity is affected by the gasproducing characteristics of the catalyst, the FCCU feed quality (particularly the sulfur content), and the reactor temperature. The FG helps to isolate the catalyst contribution in the C1-C4 fraction. Hydrogen production is measured to complete the picture of catalytic effects upon total gas yield. Traditionally, H2 yields have been reported using the H2/CH4 ratio, and we also use this approach. Hydrogen production is very sensitive to catalyst metals contamination. At low catalyst metals levels (<1000 ppm Ni + V, H2/CH4 ratios of less than 1.0 are typical. At catalyst Ni + V levels of 1000-1500 ppm, which is the normal level for many FCCUs, H2/CH4 ratios of 1-2 are typical. For the case of severe metals contamination when the Ni + V levels have risen above 5000 ppm. H2/CH4 ratios of 10 or higher would be expected with typical cracking catalysts available today. Such an increase in H2 production would, at constant feed rate, increase the gas volume handled by the gas plant by 50% and thus would substantially reduce the capacity of the FCCU. PHYSICAL PROPERTIES Particle size distribution. Particle size data provide information about (1) the circulating characteristics of the catalyst, (2) the cyclone performance and (3) the attrition characteristics of the system. Good catalyst circulation is achieved in an FCCU when the circulating catalyst has a broad particle size distribution. There should be a significant amount of catalyst having a particle size <40 and a significant amount >100. If these fractions become too small, circulating problems will occur. Each unit has its own characteristics. Some units run well with 3-5% in the <40 fraction while other units required 10-15% in this fraction.

As with other data on the equilibrium catalyst report sheet, it is important to watch for trends. If the 0-10 fraction shows signs of decreasing, the FCCU operator should be alert for indications of circulating problems and be prepared to take corrective action. Corrective action could include recycle of fines if they are available from an electrostatic precipitator or from external third stage cyclones or it could mean that the catalyst manufacturer is requested to supply a finer particle size grade containing a higher content of 0-40 material. Changes in equilibrium catalyst particle size can be due to a variety of causes. A decrease in fines content with a shift to a coarser average particle size could represent a loss in cyclone efficiency. An increase in fines content would probably be due to increased attrition of the catalyst, which could result from internal problems. Changes in air-flow patterns in the regenerator, for example, due to problems with the air grid, could cause increased attrition. Changes in particle size could also be due to changes in the particle size of the fresh catalyst shipments. It is thus necessary to have information on particle size of the fresh catalyst from the manufacturer to complete the analysis. Surface area, pore volume, and bulk density physical properties. When the FCCU is operating normally, these values are mainly influenced by the type of catalyst being used. In general there are two classes of catalysts available from European catalyst manufacturers: (1) a low surface area grade having a low pore volume and high bulk density, and (2) a high surface area grade having a high pore volume and a low bulk density. Big difference exists in these properties between these two grades when the fresh catalysts are evaluated. The equilibrium catalysts directionally show the same difference but the degree of difference is reduced. With equilibrium catalysts, the main differences between the physical properties of these two grades is in the surface area, with only slight differences remaining between the pore volume and bulk density of these two grades. The sieve content of the fresh catalyst also strongly influences these properties both in the fresh and equilibrium forms, particularly with regards to surface area. Increasing sieve content significantly increases surface area while slightly increasing pore volume and slightly decreasing bulk density. To illustrate these effects, a typical comparison between these two catalyst categories for a 10-11% sieve catalyst is shown in Table 3. For fresh catalyst, these properties are measured after calcining the sample at 400C for one hour (surface area) or at 340C for 3 hours (pore volume and bulk density). The equilibrium catalysts are tested in as-received condition.

For abnormal regenerator operations resulting in a big decrease in MAT activity, the surface area, pore volume and bulk density data of the equilibrium catalyst provide some insight into the cause of the problem. As mentioned earlier, severe catalyst deactivation can be caused by an upset in the regenerator. Hydrothermal catalyst deactivation causes a destruction of the molecular sieve component of the catalyst and hence a loss of sieve surface area. Typically the decrease in surface area in M2/GM is about twice the drop in MAT values. As mentioned earlier, the matrix component of equilibrium catalyst is resistant to hydrothermal deactivation. Since the equilibrium catalyst pore volume and bulk density are mainly associated with the matrix, these properties do not change when the catalysts is hydrothermally deactivated.
TABLE 3 Comparing surface area effects
Fresh catalyst Type, surface area Surface area, m /g Pore volume, cc/g Bulk density, g/cc
2

Equilibrium catalyst Low 56 0.27 0.82 High 110 0.35 0.76

Low 100 0.24 0.68

High 300 0.42 0.53

TABLE 4 Effect of deactivation type

Hydrothermal Surface area Pore volume Bulk density Decreases Slight change Slight change Thermal Decreases Decreases Increases

TABLE 5 Effect of hydrothermal deactivation

Before Activity, MAT
2 Surface area, m /g

After 60 56 0.29 0.79

66 72 0.30 0.79

Pore volume, cc/g Bulk density, g/cc

At extremely high temperatures (>1000C) matrix collapse can occur along with sieve destruction. This is known as thermal deactivation. It results in loss of surface area for both the sieve and matrix components and shrinkage of the matrix component, which causes a loss of equilibrium catalyst pore volume and an increase in bulk density. Thus, when a large drop in catalyst activity is noted, a study of the equilibrium catalyst physical data can some times define the cause by the changes listed in Table 1. A typical example of the change in equilibrium catalyst properties due to hydrothermal deactivation is shown in Table 5.

CHEMICAL ANALYSIS The equilibrium catalyst data sheet shows the amount of the most common contaminants found on cracking catalysts nickel, copper, vanadium, iron, and sodium. Metals. As mentioned earlier, the deposition of contaminant metals on the catalyst causes a loss of catalyst activity and selectivity. Nickel, copper, vanadium, and iron are all catalysts themselves. They are particularly effective in catalyzing hydrogenation and dehydrogenation reactions and they have some cracking activity. Such reactions are also catalyzed in the cat cracker by these metals, resulting in an increase in hydrogen, olefins, coke, and light hydrocarbons with a decrease in gasoline yield. Nickel, copper, and vanadium come from high molecular weight organo-metallic compounds in the cracker feed. It has been reported5 that essentially all of these metals in the feed deposit on the catalyst. Thus, if the feed contains 0.2 ppm of nickel and the catalyst makeup rate is 0.5 kg/ton of fresh feed (0.15 lbs/bbl)., then the equilibrium catalyst would be expected to contain 400 ppm of nickel. Only a portion of the iron is believed to come in with the feed as an organo-iron compound. This organo-iron compound, when deposited upon the equilibrium catalyst, is catalytically active. The major portion of the iron on the catalyst is believed to have originated from metal scale from the equipment and from the fresh catalyst itself. This iron is not catalytically active. For catalysts containing up to 2000 ppm of Ni + V, it has been reported3 that nickel has a much stronger effect than vanadium. Copper is believed to be equivalent in reactivity to nickel. The deposited iron originating as an organic compound is also believed to be similar in activity to nickel and copper. However, since most of the iron on the catalyst came from an inert source, the total iron on the catalyst is considered to be much less active. The total metals on the catalyst are frequently reported as ppm of equivalent nickel. A common formula for equivalent nickel is: Equiv. Ni, ppm = Ni + Cu + V/5 + Fe/10 with all values given in ppm. The large difference in reactivity of nickel and vanadium is possibly due to the manner in which these two metals are deposited upon the equilibrium catalyst. It has been reported6 that vanadium accumulates at the surface of the particle while nickel penetrates uniformly throughout the catalyst particle. Recent studies at high metals levels (>5000 ppm) have indicated that the relative activity of nickel and vanadium may be reversed. At these high levels sodium may be a much more active poison than nickel.

Sodium. This severe catalyst contaminant usually exists as sodium chloride in the feed, resulting from a combination of poor desalter and poor vacuum-tower operations. Only a portion of the sodium in the cracker feed (<20%) normally is deposited on the catalyst. Sodium levels in the feed of 1-2 ppm are normal and present no problems to the catalyst. The effect of sodium upon catalyst activity is well known. Not much has been written about other possible side effects of sodium poisoning. Conceivably, the loss of hydrogen transfer capability accompanying the loss of sieve could result in poorer gasoline selectivity. In the sodium-poisoning case that was mentioned earlier, there was a definite loss of gasoline selectivity during the sodium-poisoning period. In this same case a surprising loss of octane was also noted that coincided exactly with the sodium poisoning period. The apparent relationship between sodium poisoning and octane loss is significant. Here an increase of 0.35 wt.% in Na2O content apparently resulted in a 1.0 number research octane loss. This surprising relationship needs additional study to determine if it is real or whether some other unmonitored variable was the real cause. Carbon. The final item on the report sheet is the carbon level of the regenerated catalyst (CRC). This I an extremely important value, so important that nearly all refineries have their own monitor of this value on a daily basis. The value on the equilibrium catalyst sheet thus provides a periodic check upon the value measured in the refinery. The CRC number is an important guide to the operations of the regenerator. When a CO combustion promoter is used, the CRC value is generally in the region 0.05-0.15 wt.%. When no promoter is used, the CRC is normally in the range 0.25-0.40 wt.%. The presence of coke on the catalyst means that some of the active zeolite sites in the catalyst are temporarily deactivated. The higher the coke level, the lower is the activity of the catalyst in the unit. It has been reported2 that a 0.10 wt.% increase in CRC will typically result in a 3% decrease in unit conversion. It should be remembered that the MAT activity values are determined after the coke on the catalyst sample has been burned off. The activity that the unit sees is the reported MAT activity minus the activity loss due to the presence of coke on the catalyst as it leaves the regenerator.